Re: frameworks, dynamic_cast and gcc 4.0
Re: frameworks, dynamic_cast and gcc 4.0
- Subject: Re: frameworks, dynamic_cast and gcc 4.0
- From: Andreas Grosam <email@hidden>
- Date: Tue, 23 Aug 2005 17:47:46 +0200
On 23.08.2005, at 13:39, Steve Baxter wrote:
Actually I think I have the answer to this. We need a copy of
libstdc++ compiled with __GXX_MERGED_TYPEINFO_NAMES defined to 0.
The GCC runtime appears to assume that if weak symbols are supported
then it is safe to compare type_info by address, as the linker will
always merge symbols. This may not be true in all cases however - if
the object is defined in more than one DSO (e.g. because it is a
template class) then each DSO will have its own type_info structure
and dynamic_cast will break.
A simple fix for this would be to compile a static version of
libstdc++ with this switch set. The problem however is that currently
Apple force use of the dynamic (system) version of libstdc++ - is
there any way to switch back to a static version?
I would not recommend to switch to a static version. The reason is,
that you need a shared version when you have to throw and catch
exceptions accross DSOs.
Cheers,
Steve.
On 22 Aug 2005, at 16:56, Steve Baxter wrote:
Hi Stephan,
I ran into a problem like this as well. The problem seems to be a
difference in implementation between dynamic_cast in GCC 4.0 and in
most other compilers (or at least CodeWarrior and VC++).
It seems that GCC 4.0 treats classes linked in different shared units
as different, even if they have the same name. There is some good
information here:
http://gcc.gnu.org/faq.html#dso
You will see this problem if the same source file is linked or
included in more than one DSO (dynamic shared object). More
irritatingly, you can run into this problem with fully-inline classes
- if the class header is included in more than one DSO, it will end
up with two type_info copies, one in each DSO (this can happen very
commonly with template code).
These constructs are said to have "vague linkage":
http://developer.apple.com/documentation/DeveloperTools/gcc-4.0.0/
gcc/Vague-Linkage.html
The reason for the problem seems to be that the GCC writers changed
GCC so that type_info was checked for equality by comparing addresses
rather than really comparing for equality. Two type_infos with
different addresses are always considered to be different even if
they are in fact identical. This was done for performance reasons -
a pointer comparison is (obviously) faster than a string compare.
I suspect this works OK within a single DSO - the linker discards
duplicate copies of the type_info constructs so you end up with only
one. I am not sure what you are supposed to do in multi-DSO
applications where the same template or inline class might be used in
several DSOs - I would want then to be treated as identical, but the
GCC runtime would not do this. This is probably going to be a
show-stopper for us - this works fine in VC++ and CodeWarrior though.
Does anyone know if there is a compiler switch or something to turn
type_info comparisons by equality rather than by address back on?
Cheers,
Steve.
On 22 Aug 2005, at 13:42, Stephan Huber wrote:
Hi all,
I switched with my code to XCode 2.1 and gcc 4.0. Now most of my
dynamic_casts are failing (they worked fine before XCode 2.1). And I
don't know why, my knowledge of the inners of the compiler etc is
limited.
To be precise: the dynamic_casts are working in the scope of the
application, but If I call a method/function inside my custom
framework which does a dynamic_cast, this dynamic_cast fails
To be more specific:
I declare a class inside my framework
class Foo : public osg::Object {...}
I create some objects inside the application and attach them to
other objects (they have a property osg::Object (which is declared
in another framework)) After a while I need to process the objects
inside my custom framework and there the dynamic_cast to Foo fails.
I found the following message in the archives
<http://lists.apple.com/archives/Xcode-users/2005/May/msg00315.html>
which describes my problem in a more complicated way, doing some
searches in the archives reveils some posts about similar problems,
the solution was to *export* the class-definition, but I can't find
any information how to do this, but maybe I'm dumb ;)
P.S. My code worked very well and reliable under gc 3.x and
Panther...
Any help is greatly appreciated,
thanks in advance,
Stephan
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
email@hidden
This email sent to email@hidden
Stephen Baxter
Software Development Manager
Improvision
email@hidden
+44-2476-692229
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
email@hidden
This email sent to email@hidden
Stephen Baxter
Software Development Manager
Improvision
email@hidden
+44-2476-692229
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
email@hidden
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden